Automatization is destroying jobs: or is it?
In the photo, an Amazon robot. It was originally created by Kiva Systems after it’s creator, Mick Mountz got tired of Webvan’s logistical costs, which eventually led the company to bankruptcy. Kiva Systems was acquired by Amazon in March 2012 for $775 million, its contracts with major distribution companies like The Gap, Walgreens, Staples, Office Depot, Crate & Barrel or Saks 5th Avenue didn’t get renewed, and it was put to work exclusively for Amazon warehouses, renamed as Amazon Robotics.
The robot, about 40 centimeters high and driven by six-wheels, is designed to move easily under the shelves of warehouses and can be unscrewed a few centimeters in height to transport shelves. Each robot weighs about 145 kgs, and can move loaded shelves weighing up to 315 kilos. When the battery discharges, it automatically returns to a charging station where it plugs itself in.
Amazon recently announced it had increased its robot army by 50%: in December 2014, the company had 15,000 robots in 10 warehouses, which became 30,000 in December 2015, and then 45,000 distributed throughout 20 warehouses in December 2016.
In addition, Amazon has launched its cash register-less Amazon Go stores, where customers use an app to identify themselves, and are followed by a set of cameras and sensors that use artificial vision, sensor fusion and deep learning to identify when somebody takes a product from a shelf or returns it to it. According to the US labor force surveys (2014), store and supermarket lines of cash registers provided employment to about three and a half million people in the United States, who earned an average annual salary of $19,310 per year ($9.28 per hour) with an estimated annual growth in employment generation of 2% per annum, below the global average.
What do we expect of a company that is rapidly increasing the number of robots in its warehouses, that eliminates tellers from its stores and that delivers goods to its customers using autonomous drones instead of using a traditional logistics? One answer would be a reduction in the number of people it employs. But no: Amazon has just announced its intention to incorporate about one hundred thousand people full time over the next 18 months in logistics and technology positions, joining the wave of companies announcing massive expansion plans and creating jobs so as to avoid, in part, harassment by the Trump administration. The company will increase its workforce from 180,000 to 280,000 workers in the United States. In 2011, Amazon employed about 30,000 people. In contrast, it has been traditional distribution companies that have steadily laid off workers over the last few years.
The process is taking place in other sectors: in the San Francisco Bay Area, taxi drivers have been losing their jobs as companies like Uber or Lyft have become the preferred transport alternative for more and more people. However, these companies are now among the largest generators of employment in the area — Uber alone employs more than twenty thousand drivers in and around San Francisco, far more than existed before its arrival, and that is without taking into account the additional employment generated by these companies in management or technology positions. Eventually, taxis will become autonomous vehicles, but for the meantime, there has been little net destruction of jobs.
What is the reason that a company like Amazon, which has come to represent robotization is generates jobs at these levels? Actually, there is no paradox here, but rather an effect that some researchers like the incredibly talented Jeremy Rifkin, whom I recently had the opportunity to see in Detroit, have noted: that automation does not immediately generate net destruction of jobs but an increase, as it becomes necessary to adapt more and more processes and structures to automatization. Thus, a development such as the move from centralized generation of electric power to a distributed system requires the work of thousands of people to adapt a country’s households to the requirements of insulation and installation of solar panels, in the same way as connecting our roads for the deployment of autonomous vehicles need an army of workers.
One more element to consider when estimating the the macro-level impact of technology: in the coming years, tasks we have always seen humans carry out will be done by robots, but that does not mean that the technology required will not generate jobs that cannot be automated for humans. In other words, job destruction is not so immediate or evident, and certainly doesn’t justify anti-technology policies that that if adopted by some countries would only generate comparative disadvantages when adopted by others. In fact, technology-enabled and savvy companies are far less likely to destroy jobs than traditional ones.
A final point that politicians fearful of electorates worried about losing their jobs. The idea of keeping people in positions simply because they have to do something when the technology exists to carry out those tasks better and more economically is simply absurd.
The repercussions of the ongoing process of automatization are not as simple or obvious as some people would like to believe. So before drawing easy conclusions, we need to look at the facts in great detail…
(En español, aquí)
